Everything in hard money — flows, housing, and CRE priced in gold & silver (1998→2026)
Built 2026-06-07 from models/graph/gold_silver_reprice.py → data/gold_silver_reprice.json. Annual-average gold/silver (LBMA/macrotrends/USGS; 2024 gold avg $2,386.20 confirmed LBMA), US median home (FRED MSPUS), CRE (Green Street CPPI, 2007 peak=100), regional homes (S&P/Case-Shiller, Jan 2000=100). Recent years (2025-26) flagged ~approx given the metals spike.
Why re-price in gold/silver
A dollar measuring stick that is itself being debased exaggerates "gains." Denominating in ounces strips the monetary illusion and shows the real change in purchasing power. The result reframes the entire analysis: a large share of the "bubble" is currency debasement, not asset appreciation — which is exactly why gold itself is at records and central banks are buying it (commodities-metals.json).
1. The capital flows in hard money (the debasement reveal)
| Year | Flow | $ | in gold |
|---|---|---|---|
| 1998 | LTCM bailout | $3.6B | 12 M oz |
| 2008 | TARP | $700B | 803 M oz |
| 2020 | COVID Fed B/S expansion | $4.5T | 2,542 M oz |
| 2025 | OpenAI compute commitments | $1.4T | 424 M oz |
| 2026 | AI datacenter private-credit pipeline | $800B | 186 M oz |
- OpenAI's $1.4T is 389× LTCM in dollars — but only ~35× in gold. The apparent explosion of scale is mostly the shrinking dollar.
- OpenAI's commitments ≈ 0.53× the entire 2008 TARP in gold-ounce terms — enormous, but the 2020 COVID monetary expansion (2,542 M oz) dwarfs even that. The biggest "flow" of the era was monetary, not corporate.
2. US median home — the housing "gains" are mostly debasement
| Year | $ | gold-oz | $ idx (1998=100) | gold idx (1998=100) |
|---|---|---|---|---|
| 1998 | $152k | 517 | 100 | 100 |
| 2007 | $247k | 355 | 162 | 69 |
| 2012 | $245k | 147 | 161 | 28 |
| 2020 | $336k | 190 | 221 | 37 |
| 2022 | $457k | 254 | 301 | 49 |
| 2026 | $418k | 97 | 275 | 19 |
A US home is up 175% in dollars since 1998 but down ~81% in gold (and ~78% in silver: 27,387 oz → 5,971 oz). The celebrated 2020-22 boom barely registered in hard money (gold-idx 37→49) and is now falling hard (→19). Housing has been a wealth-preservation loser vs hard money for a quarter-century — the nominal gains were the currency melting, not the house appreciating.
3. Commercial real estate (Green Street CPPI) in gold
CRE peaked in hard money around 2001 (gold-idx 122) and has fallen to 16 (2026) — down ~84% in gold — even though the nominal CPPI more than doubled (55→129). The office/CRE unwind (macro-cre-privatecredit.md) is brutal in dollars; in gold it has been a 25-year secular decline. The leverage stacked on CRE (bank HTM, private credit) was lent against a nominal price that, in real money, was eroding the whole time.
4. Regional residential divergence — collapses and inverts in gold
| Metro | 2000 (nom) | 2026 (nom) | 2026 gold-idx (2000=100) |
|---|---|---|---|
| Los Angeles | 100 | 445 | 29 |
| Miami | 100 | 475 | 31 |
| Detroit | 100 | 198 | 13 |
| Chicago | 100 | 190 | 12 |
The nominal story is "coasts boomed (LA 445, Miami 475), heartland lagged (Detroit 198)." In gold, everyone lost — the coasts merely lost less (~70% down) while the heartland was devastated (~88% down). The regional divergence the user observed is real in dollars but is, in hard money, a common regional debasement with different severities — which is exactly why a single national policy rate cannot address it (macro-fed-trap-regional.md).
5. The synthesis
Re-pricing in gold/silver unifies the whole investigation under one mechanism: for 25 years, the dollar has been the shrinking ruler. The AI capex loop, the housing/CRE "wealth," and the bank-balance-sheet values are all measured in a unit that has lost ~80% of its purchasing power vs gold since 1998-2000. That is why (a) the bubble can look simultaneously enormous (dollars) and "only half a TARP" (gold); (b) gold/silver are at records with central banks accumulating and silver in backwardation; and (c) the Fed's single rate cannot fix it — the divergence is partly a monetary-denominator problem no interest rate resolves. The hard-money lens is the through-line connecting Layer 1 (AI), Layer 2 (banks/RE), and Layer 3 (metals).
← Research index · structured data: macro-gold-silver-reprice.json · macro-gold-silver-reprice.md